The long-term objective of this project is to create the technology needed to make human brain functional magnetic resonance imaging (FMRI) into an interactive clinical and experimental tool. One result of changing FMRI from a "batch" to a "real-time" orientation is that it will become much easier to carry out complex investigations of brain function and to be assured that quality data have been gathered - problems will be seen and dealt with immediately. This is especially critical for clinical applications in neurology and neurosurgery, since the clinician will know that the needed information has been gathered before the patient leaves the MR scanner (unlike the present situation). The second major implication of this work is that new clinical and research applications of FMRI will become possible when an investigator can control the stimulus presented to the patient/subject based on up-to-the-minute results. The theme of the proposed work is that more information can be extracted from FMRI data than current methods allow, that the necessary processing can be done in real-time, and that this will make FMRI into a much more useful tool for probing the human brain than it is now. The primary motivation is to greatly improve the utilization of very limited patient and scanner time. The applicants proposed to: 1) Generalize existing methods to analyze functional activation from MRI time series to deal with complex stimulus patterns; 2) Produce methods to automatically measure the quality of an FMRI time series; detect and compensate for intra-scan variations (e.g. subject motion); 3) Improve methods to compare high resolution 3-D functional activation maps; 4) Design new techniques for "on-line" computation to avoid "dead time" after the end of a scanning run. All procedures will be tested and refined by applying them to studies of the visual and motor cortices in normal volunteers. These developments are required to make interactive FMRI a useful reality. One application of the new methods that will be developed in detail is the interactive adjustment of stimulus levels of map individual sensitivities in the human visual cortex. This will allow the mapping of neurological visual field deficits while a patient is in the MR scanner. Real-time interactive FMRI has the potential for major long term health significance: regular clinical analyses of patient brain function and new avenues of research in neuroscience will both become practical.